JP2006270818A - Server - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a server with sufficient processing ability for an FEC process of a plurality of streams. <P>SOLUTION: An FEC server 1 includes a plurality of modules 11-1 to 11-n for performing the FEC process of the streams. A distribution module 13 discriminates to which stream received packets belong and distributes the packets to any one of the plurality of the FEC modules 11-1 to 11-n depending on the results discriminated. Consequently, the number of the FEC module can be easily varied, thus making it possible to provide a server with the sufficient processing ability. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a server connecting a plurality of networks, and more particularly to a server having an FEC (Forward Error Correction) encoding function or an FEC decoding function for recovering streaming packet loss.

  In recent years, the Internet has become widespread, and users can access various information on sites operated in various parts of the world and obtain the information. Among devices connected to such a network, there is a device that is connected between a plurality of networks and relays a stream.

For example, the relay device disclosed in Japanese Patent Application Laid-Open No. 2004-15551 decodes the received FEC data and increases the decoding level when the redundancy level of the received FEC data is lowered. The encoded data is re-encoded and transmitted to the downstream side.
JP 2004-15551 A

  When the redundancy level is lowered, the relay device disclosed in Japanese Patent Application Laid-Open No. 2004-15551 described above decodes the received FEC data and re-encodes the decoded data for transmission. If the redundancy level is not lowered, the received FEC data is transmitted as it is. Therefore, the downstream network is a network with little packet loss, and even when there is no need to perform FEC encoding, an FEC encoded stream is always transmitted, which causes a problem that traffic increases.

  In addition, the relay apparatus encodes and transmits a plurality of streams, but the FEC encoding processing time greatly depends on the FEC parameters. Therefore, when the relay device processes a plurality of streams having different FEC parameters, the time required for the processing changes greatly. Since the processing capability of the relay device is constant, there is a problem that the processing cannot be completed if the processing time is long, or the processing capability is wasted if the processing time is short.

  Similarly, the relay apparatus receives and decodes a plurality of streams, but the FEC decoding processing time greatly depends on the FEC parameters. Therefore, when the relay device processes a plurality of streams having different FEC parameters, the time required for the processing changes greatly. Since the processing capability of the relay device is constant, there is a problem that the processing cannot be completed if the processing time is long, or the processing capability is wasted if the processing time is short.

  The present invention has been made to solve the above problems, and a first object thereof is to provide a server capable of reducing traffic in a network with little packet loss that does not require FEC processing. .

  The second object is to provide a server having necessary and sufficient processing capability for FEC processing of a plurality of streams.

  According to one aspect of the present invention, there is provided a server that connects a network with a high packet loss and a network with a low packet loss, the receiving means for receiving a stream via the network with a low packet loss, and a receiving means. Encoding means for encoding the received stream into an error-correctable format, and transmission means for transmitting the stream encoded by the encoding means via a network with many packet losses.

  Preferably, the server includes a plurality of encoding means, and the server further determines to which stream the packet received by the receiving means belongs, and distributes the packet to any of the plurality of encoding means according to the determination result. Including sorting means.

  More preferably, the server further determines whether or not the newly registered stream matches the registered stream based on the transmission source address and the destination address of the newly registered stream. Control means for determining to which encoding means a stream to be newly registered is assigned according to the load status of the encoding means and instructing the distribution means is included.

  More preferably, the control means newly registers based on the sum of the stream rates determined for each error correction parameter and the maximum stream rate for each error correction parameter that can be processed by one encoding means. Decide which encoding means to distribute the stream to.

  According to another aspect of the present invention, a server that connects a network with a high packet loss and a network with a low packet loss, and receives a stream encoded in an error-correctable format via the network with a high packet loss. Receiving means for decoding, a decoding means for decoding the stream received by the receiving means, and a transmission means for transmitting the stream decoded by the decoding means via a network with little packet loss.

  Preferably, the server includes a plurality of decoding means, and the server further determines to which stream the packet received by the receiving means belongs, and distributes the packet to one of the plurality of decoding means according to the determination result. Including sorting means.

  More preferably, the server further determines whether or not the newly registered stream matches the registered stream based on the transmission source address and the destination address of the newly registered stream. Control means for determining to which decoding means a stream to be newly registered is assigned and instructing the distribution means in accordance with the load status of the decoding means is included.

  More preferably, the control means newly registers based on the sum of the stream rates determined for each error correction parameter and the maximum stream rate for each error correction parameter that can be processed by one decoding means. It is determined to which decoding means the stream is distributed.

  According to yet another aspect of the present invention, a server that connects a network with a high packet loss and a network with a low packet loss and includes a master server and a slave server, the master server having a packet loss First receiving means for receiving a stream via a network with a low network or an adjacent slave server, and a first plurality for encoding the stream received by the first receiving means into an error-correctable format To determine which stream the packet received by the encoding means and the first receiving means belongs to, and distribute the packet to either the slave server or the first plurality of encoding means according to the determination result The first encoded means and the first encoded means encoded by the plurality of encoded means. A first transmission means for transmitting a network via a network with a high packet loss or an adjacent slave server, wherein the slave server is a network with a low packet loss, a master server or another adjacent slave. A second receiving means for receiving the stream via the server, a second plurality of encoding means for encoding the stream received by the second receiving means into an error-correctable format, A second distribution unit for determining which stream the packet received by the reception unit belongs to and distributing the packet to one of the second plurality of encoding units according to the determination result; The stream encoded by other encoding means is used for the network and master server where there is a lot of packet loss. Or a second transmission means for transmitting via the adjacent slave servers.

  According to yet another aspect of the present invention, a server that connects a network with a high packet loss and a network with a low packet loss and includes a master server and a slave server, the master server having a packet loss First receiving means for receiving a stream encoded in an error-correctable format via a network with many networks or adjacent slave servers, and a first receiving means for decoding the stream received by the first receiving means A plurality of decoding means and a stream to which the packet received by the first receiving means belongs, and the packet is sent to either the slave server or the first plurality of decoding means according to the determination result Decoded by the first sorting means for sorting and the first plurality of decoding means First transmission means for transmitting the stream via a network with low packet loss or a server of an adjacent slave, wherein the slave server is connected to a network with low packet loss, a master server, or another adjacent slave. Second receiving means for receiving a stream encoded in an error-correctable format via a server, and a plurality of second decoding means for decoding the stream received by the second receiving means; A second distribution unit for determining which stream the packet received by the second reception unit belongs to, and distributing the packet to any one of the second plurality of decoding units according to the determination result; Streams decoded by multiple decoding means in a network with little packet loss And a second transmission means for transmitting through the master server or adjacent slave server.

  According to an aspect of the present invention, the encoding unit encodes the stream received by the receiving unit into an error-correctable format, and the transmitting unit transmits the stream encoded by the encoding unit via a network having a large packet loss. Since transmission is performed, it is possible to reduce traffic in a network with little packet loss.

  In addition, since the distribution unit determines which stream the packet received by the reception unit belongs to, and distributes the packet to one of a plurality of encoding units according to the determination result, the number of encoding units can be easily changed. Therefore, it is possible to provide a server having necessary and sufficient processing capacity.

  Further, the control means determines whether or not the newly registered stream matches the registered stream based on the transmission source address and the destination address of the newly registered stream. Depending on the load status, the encoding method is determined and the distribution unit is instructed to allocate the newly registered stream, so that the encoding process cannot be completed or the processing capacity of the server is wasted. Can be prevented.

  In addition, the control unit determines a stream to be newly registered based on the sum of the stream rates determined for each error correction parameter and the maximum stream rate for each error correction parameter that can be processed by one encoding unit. Since it is determined which encoding means to distribute, it becomes possible to easily determine the load status of the plurality of encoding means.

  According to another aspect of the present invention, the decoding unit decodes the stream received by the receiving unit, and the transmitting unit transmits the stream decoded by the decoding unit via a network with little packet loss. It becomes possible to reduce traffic in a network with little loss.

  In addition, since the distribution unit determines which stream the packet received by the reception unit belongs to, and distributes the packet to one of a plurality of decoding units according to the determination result, the number of decoding units can be easily changed. Therefore, it is possible to provide a server having necessary and sufficient processing capacity.

  Further, the control means determines whether or not the newly registered stream matches the registered stream based on the transmission source address and the destination address of the newly registered stream. Depending on the load status of the system, the decoding unit to which the newly registered stream is to be allocated is determined and the distribution unit is instructed, so that the decoding process cannot be completed or the processing capacity of the server is wasted. Can be prevented.

  In addition, the control means determines the stream to be newly registered based on the sum of the stream rates determined for each error correction parameter and the maximum stream rate for each error correction parameter that can be processed by one decoding means. Since it is determined which of the decoding means is assigned, it is possible to easily determine the load status of the plurality of decoding means.

  According to still another aspect of the present invention, the first distribution unit determines to which stream the packet received by the first reception unit belongs, and the packet is sent to the slave server or the second server according to the determination result. Since it is distributed to any one of a plurality of encoding means, the number of slave servers and the number of encoding means can be easily changed, and a server having necessary and sufficient processing capability can be provided. .

  According to still another aspect of the present invention, the first distribution unit determines to which stream the packet received by the first reception unit belongs, and the packet is sent to the slave server or the second server according to the determination result. Since it is distributed to any one of the plurality of decoding means, the number of slave servers and the number of decoding means can be easily changed, and it becomes possible to provide a server having necessary and sufficient processing capability. .

(First embodiment)
FIG. 1 is a diagram illustrating a configuration example of a system including an FEC server according to the first embodiment of this invention. The FEC server 1 is connected to a network with a large packet loss, such as the Internet 2, and a network with a small packet loss, such as an intranet 3, and has an FEC encoding function and an FEC decoding function.

Here, a network with a lot of packet loss refers to a network in which packet loss occurs at a rate of about 10 ⁻³ to 10 ⁻¹ . For example, in a 4 Mbps stream, if one packet is 1000 bytes, the number of packets per second is 500, but if the packet loss rate is 10 ⁻³ , packet loss occurs once every 2 seconds. Become.

A network with little packet loss refers to a network in which packet loss occurs at a rate of 10 ⁻⁷ or less. If the packet loss rate is 10 ⁻⁷ , a packet loss occurs once every approximately 5.5 hours. Needless to say, the definitions of a network with a large packet loss and a network with a small packet loss are not limited to these values.

  FEC methods applicable to the FEC server 1 include Raptor code, LT (Luby Transform) code, LDPC (Low Density Parity Check) code, Hamming code, BCH (Bose Chaudhuri-Hocquenghem) code, Reed-Solomon code, algebraic geometry There are error correction codes such as codes.

  A streaming server 4 such as a monitoring camera and a streaming client 5 such as an STB (Set Top Box) are connected to the Internet 2. The streaming server 4 has an FEC encoding function, and the streaming client 5 has an FEC decoding function.

  In addition, a streaming client 6 such as a storage server that does not have an FEC function and a streaming server 7 such as a distribution server that does not have an FEC function are connected to the intranet 3. Existing streaming clients 6 and streaming servers 7 can be used.

  When a stream is transmitted from the streaming server 4 on the Internet 2 to the streaming client 6 on the intranet 3, the FEC server 1 performs FEC decoding on the FEC-encoded stream received from the streaming server 4, and generates a normal stream (FEC encoded). Stream)) to the streaming client 6.

  When a stream is transmitted from the streaming server 7 on the intranet 3 to the streaming client 5 on the Internet 2, the FEC server 1 performs FEC encoding on a normal stream (non-FEC encoded stream) received from the streaming server 7. The FEC encoded stream is transmitted to the streaming client 5.

  FIG. 2 is a diagram illustrating a hardware configuration of the FEC server 1. The FEC server 1 is connected to a plurality of FEC modules 11-1 to 11-n that perform FEC processing, a control module 12 that performs overall control of the FEC server 1, an Internet port 14, and an intranet port 15. And an Ethernet (registered trademark) switch 13 that distributes the plurality of FEC modules 11-1 to 11-n.

  The FEC modules 11-1 to 11-n are detachable, and an arbitrary number of FEC modules can be mounted according to use conditions. Each of the FEC modules 11-1 to 11-n includes a CPU (Central Processing Unit) 21 and a memory 22. When the CPU 21 executes a program stored in the memory 22, FEC processing (FEC encoding, FEC) Decode). A plurality of FEC modules 11-1 to 11-n execute their respective programs, thereby realizing distributed processing.

  The control module 12 includes a CPU 31 and a memory 32, and performs overall control of the FEC server 1 when the CPU 31 executes a program stored in the memory 32.

  The Ethernet (registered trademark) switch 13 is connected to the Internet port 14, the intranet port 15, the FEC modules 11-1 to 11-n and the control module 12. The Ethernet (registered trademark) switch 13 distributes the FEC-encoded stream received via the Internet port 14 to the FEC modules 11-1 to 11-n. Then, the normal stream after being decoded from the FEC modules 11-1 to 11-n is received, and the stream is transmitted via the intranet port 15.

  Further, the Ethernet (registered trademark) switch 13 distributes the normal stream received via the intranet port 15 to the FEC modules 11-1 to 11-n. Then, FEC encoded streams are received from the FEC modules 11-1 to 11-n and transmitted via the Internet port 14.

  Each of the Internet port 14 and the intranet port 15 has a performance of 1 Gbps Ethernet (registered trademark). The interface between the Ethernet (registered trademark) switch 13 and the FEC modules 11-1 to 11-n and the control module 12 has a performance of 100 Mbps Ethernet (registered trademark).

  FIG. 3 is a diagram schematically illustrating the FEC server 1 transmitting the stream received through the intranet port 15 after FEC encoding. When the distribution module 13 receives a normal stream via the intranet port 15, the distribution module 13 sends the stream to the FEC module to be processed. Then, the FEC encoded stream is transmitted to the streaming client 5 through the Internet port 14.

  FIG. 4 is a diagram schematically showing the FEC server 1 transmitting the FEC-encoded stream received via the Internet port 14 by FEC decoding. When the distribution module 13 receives the FEC-encoded stream via the Internet port 14, the distribution module 13 sends the stream to the FEC module to be processed. Then, the FEC decoded stream is transmitted to the streaming client 6 via the intranet port 15.

  The distribution module 13 is provided on each of the Internet port 14 side and the intranet port 15 side, and the Ethernet (registered trademark) switch 13 has these functions.

  The control module 12 sets the session information and FEC parameters of the stream by setting from a management device (not shown) or snooping a session setting protocol such as RTSP (Real-Time Streaming Protocol) between the streaming client and the streaming server. get.

  As will be described later, the control module 12 determines which FEC module to process the stream according to the FEC parameter of the stream and the load status of the FEC module. Then, the control module 12 sets the session information of the stream and the FEC module for processing in the distribution module 13, and sets the session information of the stream and the FEC parameter in the FEC module for processing.

  The session information includes an IP (Internet Protocol) address between the streaming server and the streaming client, a UDP (User Datagram-Protocol) port number of the streaming client, a stream rate, and the like.

  The FEC parameters include protocol version, source block size, encoding unit size, target loss rate, network loss rate, designation of FEC encoding or FEC decoding, and the like.

  Based on the session information received from the control module 12, the distribution module 13 determines whether or not a packet received via the Internet port 14 or the intranet port 15 belongs to a stream to be subjected to FEC processing. When it is determined that the received packet belongs to a stream to be subjected to FEC processing, the packet is sent to the FEC module to process the stream. If it is determined that the received packet does not belong to the stream to be subjected to FEC processing, the packet is transmitted as it is from the other port.

  If the source IP address of the received packet matches the IP address of the streaming server, and the destination IP address of the received packet matches the IP address of the streaming client, the distribution module 13 is a stream to which the packet is to be subjected to FEC processing. It is determined that it belongs to.

  The FEC modules 11-1 to 11-n perform the FEC process only when the destination UDP port number of the received packet matches the UDP port number of the streaming client among the received packets received from the distribution module 13. The packet is not subjected to FEC processing. Then, the FEC modules 11-1 to 11-n change the IP Identification field, change the checksum, etc. for all received packets. The generated packet is transmitted from the other port.

  Next, a method for the control module 12 to determine the FEC module that should process the stream will be described. The number of streams that can be processed by one FEC module is the number of streams that satisfies the condition that the sum of the processing capacities required for processing individual streams is equal to or less than the processing capability of one FEC module.

  The processing power required to process individual streams depends on the FEC parameters and the stream rate. The relationship between required processing power and FEC parameters cannot be expressed simply, but the required processing power is proportional to the stream rate. Therefore, the condition for determining the number of streams that can be processed by one FEC module can be expressed as follows.

  Here, ei is the sum of the rates of the FEC parameter i stream (FEC encoding), di is the sum of the rates of the FEC parameter i stream (FEC decoding), and Ei is the stream of the FEC parameter i that can be processed by one FEC module. The maximum rate of (FEC encoding), Di indicates the maximum rate of a stream of FEC parameter i (FEC decoding) that can be processed by one FEC module.

  The stream rate is the original rate of the stream, that is, the rate before FEC encoding in the case of FEC encoding, and the rate after FEC decoding in the case of FEC decoding.

  There is no simple method for determining Ei and Di, but these values are basically determined by the processing capacity of the CPU. Therefore, these values are measured in advance and used. The control module 12 actually measures the rate of each stream, and ei and di are determined based on the values.

  FIG. 5 is a diagram illustrating an example of a table that stores values of Ei and Di. As shown in FIG. 5, for each FEC parameter, the maximum rates Ei and Di that can be processed by one FEC module are held. For example, the table indicated by the index N holds the maximum rates En and Dn of the stream of the FEC parameter n that can be processed by one FEC module.

  This table also holds default values for Ei and Di. These default values are used for streams whose FEC parameters do not match any of the other tables and can be optionally set. If this default value is not set, processing of streams that do not match any of them is not performed.

  When a stream is added to the FEC module, the control module 12 adds the stream to the FEC module having the smallest value on the right side of the conditional expression shown in Expression (1) so that the load on the FEC module is equalized. However, when there are a plurality of streams between the same streaming server and streaming client, that is, when there are a plurality of streams having the same IP address, these streams all need to be processed by one FEC module. Therefore, when a stream between the same streaming server and streaming client is already set, the stream is added to the same FEC module.

  FIG. 6 is a flowchart for explaining the processing procedure of the control module 12 at the time of stream registration of the FEC server 1 according to the first embodiment of the present invention. First, the control module 12 acquires session information and FEC parameters of a stream to be newly registered from a management device (not shown) or the like (S11), and acquires an index i corresponding to the FEC parameters from the table shown in FIG. S12).

  Next, the control module 12 has a stream in which the IP address of the streaming server is the same as the IP address of the streaming server to be newly registered and the IP address of the streaming client is newly registered in the registered stream. It is determined whether there is a stream that is identical to the IP address of the streaming client (S13). If such a stream has already been registered (S13, Yes), the FEC module j to which the stream is registered is provisionally determined as a new stream distribution destination (S14), and the process proceeds to step S17.

  If such a stream is not registered (S13, No), the control module 12 calculates the value of the above formula (1) regarding the registered stream having the FEC parameter i of each FEC module (S15). The FEC module j having the smallest calculation result is provisionally determined as a new stream distribution destination (S16), and the process proceeds to step S17.

  In step S17, the rate of the stream to be newly registered is added, and the value of the above equation (1) regarding the stream having the FEC parameter i of the FEC module j is calculated.

  When the calculation result is 1 or less (S18, 1 or less), the control module 12 registers the session information of the stream to be newly registered in the internal management table (not shown), the FEC parameter, and the index j of the distribution destination FEC module ( S19). Then, session information and FEC parameters of a stream to be newly registered are registered in the FEC module j (S20).

  Then, the session information of the stream to be newly registered and the index j of the distribution destination FEC module are registered in the distribution module 13 (S21), and the process ends. In the case of FEC encoding, these pieces of information are registered in the distribution module 13 on the intranet port 15 side, and in the case of FEC decoding, these pieces of information are registered in the distribution module 13 on the Internet port 14 side.

  If the calculation result is greater than 1 (greater than S18, 1), the control module 12 stops registering a new stream (S22) and ends the process.

  FIG. 7 is a flowchart for explaining a processing procedure of the FEC module 11 at the time of stream registration of the FEC server 1 according to the first embodiment of this invention. First, the FEC module 11 acquires session information and FEC parameters of a stream to be newly registered from the control module 12 (S31). Then, session information and FEC parameters of a stream to be newly registered are registered in an internal management table (not shown) (S32), and the process ends.

  FIG. 8 is a flowchart for explaining a processing procedure of the distribution module 13 at the time of stream registration of the FEC server 1 according to the first embodiment of this invention. First, the distribution module 13 acquires the session information of the stream to be newly registered and the index j of the distribution destination FEC module from the control module 12 (S41). Then, the session information of the stream to be newly registered and the index j of the distribution destination FEC module are registered in an internal management table (not shown) (S42), and the process ends.

  FIG. 9 is a flowchart for explaining the processing procedure of the distribution module 13 during the FEC processing of the FEC server 1 according to the first embodiment of this invention. First, when the distribution module 13 receives a packet via the Internet port 14 or the intranet port 15 (S51), the IP address of the streaming server is the same as the source IP address of the received packet in the registered stream. In addition, it is determined whether there is a stream in which the IP address of the streaming client is the same as the destination IP address of the received packet (S52).

  If such a stream has already been registered (S52, Yes), the received packet is transferred to the registered distribution destination FEC module j (S53), and the process ends. If such a stream is not registered (S52, No), the received packet is transmitted from the port opposite to the receiving side, that is, the Internet port 14 in the case of FEC encoding and the intranet port 15 in the case of FEC decoding. (S54), and the process ends.

  FIG. 10 is a flowchart for explaining the processing procedure of the FEC module 11 during the FEC processing of the FEC server 1 according to the first embodiment of this invention. First, when the FEC module 11 receives a packet from the distribution module 13 (S61), it is determined whether or not there is a stream in the registered stream in which the UDP port number of the streaming client is the same as the destination UDP port number of the received packet. Is determined (S62).

  If such a stream has already been registered (S62, Yes), the FEC process of the stream is performed. That is, when the received packet is divided by the IP fragment, data is assembled (S63), FEC processing is performed on the received packet according to the registered session information and FEC parameters, and a transmitted packet is generated ( S64). At this time, if the data does not fit in one packet, the transmission packet is divided by the IP fragment (S65), and the process proceeds to step S67.

  If such a stream is not registered (S62, No), the received packet is used as it is as a transmission packet without performing the process on the stream (S66), and the process proceeds to step S67.

  In step S67, the FEC module 11 changes the Identification field of the IP header of the transmission packet to be unique among packets having the same source IP address and destination IP address, and the value of the IP header checksum. Is recalculated and stored in the transmission packet (S67).

  Then, the FEC module 11 transfers the transmission packet via the port opposite to the receiving side, that is, the Internet port 14 in the case of FEC encoding, and the intranet port 15 in the case of FEC decoding (S68).

  As described above, according to the FEC server in the present embodiment, FEC-encoded streams received from a network with a lot of packet loss such as the Internet 2 are FEC-decoded, and the packet loss of the intranet 3 or the like is reduced as a normal stream. Since the data is transferred to a small number of networks, it is possible to reduce traffic in a network with little packet loss that does not require FEC processing.

  In addition, since the streams are distributed to the plurality of FEC modules based on the above formula (1), it is possible to prevent the FEC processing from being completed or the processing capacity of the server from being wasted. It was.

  In addition, since the number of FEC modules installed in the FEC server 1 can be changed according to the transmission capability of the distribution server, the reception capability of the storage server, the FEC parameters of individual streams, etc., it has necessary and sufficient processing capability. It became possible to build a FEC server.

(Second Embodiment)
The FEC server according to the second embodiment of the present invention is configured such that a plurality of FEC servers described in the first embodiment are cascade-connected to perform distributed processing. Note that the hardware configuration of the FEC server in the present embodiment is the same as the hardware configuration of the FEC server in the first embodiment shown in FIG. 2, and only the program stored in the control module is different. Therefore, detailed description of overlapping configurations and functions will not be repeated.

  FIG. 11 is a diagram schematically illustrating the FEC server transmitting the stream received via the intranet port 15 after FEC encoding according to the second embodiment of the present invention. This FEC server has a configuration in which a plurality of FEC servers described in the first embodiment are cascade-connected, one of which functions as a master FEC server, and the other is a slave FEC server. Function as. The control module 12-m of the master FEC server 1-m controls the entire FEC server.

  When the distribution module 13 receives a normal stream via the intranet port 15, if the stream is to be processed by its own FEC server, the distribution module 13 sends the stream to the FEC module to be processed. The FEC-encoded stream is transmitted via the Internet port 14. If the stream is to be processed by another FEC server, the stream is transferred to the FEC server.

  The control module 12-m of the master FEC server 1-m performs stream session information and FEC by setting from a management device (not shown) or snooping of a session setting protocol such as RTSP between the streaming client and the streaming server. Get parameters.

  The control module 12-m determines which FEC module of which FEC server should process the stream according to the FEC parameter of the stream and the load status of the FEC module. Then, if the FEC module belongs to its own FEC server 1-m, the control module 12-m sets the session information of the stream and the FEC module for processing in the distribution module 13, and performs the processing. Session information and FEC parameters of the stream are set in the module.

  In addition, if the FEC module belongs to another FEC server, the control module 12-m sets the session information, FEC parameters, and the FEC module for processing in the control module of the FEC server. When receiving the information from the master FEC server 1-m, the control module of the slave FEC server performs the processing described in the first embodiment, thereby controlling the entire slave FEC server.

  FIG. 12 is a diagram schematically showing the FEC-encoded stream received by the FEC server via the Internet port 14 after being FEC-decoded and transmitted in the second embodiment of the present invention. When the distribution module 13 receives an FEC-encoded stream via the Internet port 14, if the stream is to be processed by its own FEC server, the distribution module 13 sends the stream to the FEC module to be processed. The FEC decoded stream is transmitted via the intranet port 15. If the stream is to be processed by another FEC server, the stream is transferred to the FEC server.

  FIG. 13 is a flowchart for explaining the processing procedure of the control module 12-m at the time of stream registration of the master FEC server 1-m in the second embodiment of the present invention. First, the control module 12-m acquires session information and FEC parameters of a stream to be newly registered from a management device (not shown) or the like (S71), and acquires an index i corresponding to the FEC parameters from the table shown in FIG. (S72).

  Next, the control module 12-m registers, in the registered stream, the IP address of the streaming server is the same as the IP address of the streaming server of the stream to be newly registered, and the IP address of the streaming client is newly registered. It is determined whether or not there is a stream that is the same as the IP address of the streaming client of the stream to be performed (S73). If such a stream has already been registered (S73, Yes), the FEC module j of the FEC server k in which the stream is registered is provisionally determined as a new stream distribution destination (S74). Proceed to processing.

  If such a stream is not registered (S73, No), the control module 12-m determines the value of the above formula (1) regarding the registered stream having the FEC parameter i of each FEC module of each FEC server. (S75), the FEC module j of the FEC server k having the smallest calculation result is provisionally determined as a new stream distribution destination (S76), and the process proceeds to step S77.

  In step S77, the rate of the stream to be newly registered is added, and the value of the above equation (1) regarding the stream having the FEC parameter i of the FEC module j of the FEC server k is calculated.

  If the calculation result is 1 or less (S78, 1 or less), the control module 12-m sends the session information, FEC parameters, index K and FEC of the distribution destination FEC server to be newly registered in the internal management table (not shown). The module index j is registered (S79). Then, the control module 12-m determines whether or not the distribution destination FEC server is its own FEC server (S80).

  If it is not its own FEC server (S80, No), the session information of the stream to be newly registered, the FEC parameter, and the index j of the distribution destination FEC module are registered in the control module 12-k of the FEC server k (S81). Exit.

  In the case of the own FEC server (S80, Yes), the session information and FEC parameters of the stream to be newly registered are registered in the FEC module j in the own FEC server 1-m (S82).

  Then, the session information of the stream to be newly registered and the index j of the distribution destination FEC module are registered in the distribution module 13 in its own FEC server 1-m (S83), and the process is terminated. In the case of FEC encoding, these pieces of information are registered in the distribution module 13 on the intranet port 15 side, and in the case of FEC decoding, these pieces of information are registered in the distribution module 13 on the Internet port 14 side.

  If the calculation result is greater than 1 (greater than S78, 1), the control module 12-m stops registering a new stream (S84) and ends the process.

  FIG. 14 is a flowchart for explaining the processing procedure of the control module 12-k at the time of stream registration of the slave FEC server 1-k according to the second embodiment of the present invention. First, the control module 12-k acquires the session information, the FEC parameter of the stream to be newly registered, and the index j of the distribution destination FEC module from the control module 12-m of the master FEC server 1-m (S91).

  Next, the control module 12-k registers the session information of the stream to be newly registered in the internal management table (not shown), the FEC parameter, and the index j of the distribution destination FEC module (S92). Then, the session information and FEC parameters of the stream to be newly registered are registered in the FEC module j (S93).

  Then, the control module 12-k registers the session information of the stream to be newly registered and the index j of the distribution destination FEC module in the distribution module 13 (S94), and ends the process. In the case of FEC encoding, these pieces of information are registered in the distribution module 13 on the intranet port 15 side, and in the case of FEC decoding, these pieces of information are registered in the distribution module 13 on the Internet port 14 side.

  As described above, according to the FEC server in the present embodiment, a plurality of FEC servers are cascade-connected to perform the FEC processing. Therefore, in addition to the effects described in the first embodiment, it is necessary. It has become possible to more flexibly construct an FEC server with sufficient processing capability.

  The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

It is a figure which shows the structural example of the system containing the FEC server in the 1st Embodiment of this invention. 2 is a diagram illustrating a hardware configuration of an FEC server 1. FIG. FIG. 3 is a diagram schematically illustrating a place where the FEC server 1 performs FEC encoding on a stream received via an intranet port 15 and transmits the stream. FIG. 3 is a diagram schematically showing a state where the FEC server 1 FEC-decodes and transmits an FEC-encoded stream received via the Internet port 14. It is a figure which shows an example of the table which stores the value of Ei and Di. It is a flowchart for demonstrating the process sequence of the control module 12 at the time of the stream registration of the FEC server 1 in the 1st Embodiment of this invention. It is a flowchart for demonstrating the process sequence of the FEC module 11 at the time of the stream registration of the FEC server 1 in the 1st Embodiment of this invention. It is a flowchart for demonstrating the process sequence of the distribution module 13 at the time of the stream registration of the FEC server 1 in the 1st Embodiment of this invention. It is a flowchart for demonstrating the process sequence of the distribution module 13 at the time of the FEC process of the FEC server 1 in the 1st Embodiment of this invention. It is a flowchart for demonstrating the process sequence of the FEC module 11 at the time of the FEC process of the FEC server 1 in the 1st Embodiment of this invention. It is a figure which shows typically the place which the FEC server in the 2nd Embodiment of this invention FEC encodes the stream received via the intranet port 15, and transmits. It is a figure which shows typically the place which the FEC server in the 2nd Embodiment of this invention FEC-decodes the FEC-encoded stream received via the Internet port 14, and transmits. It is a flowchart for demonstrating the process sequence of control module 12-m at the time of the stream registration of the master FEC server 1-m in the 2nd Embodiment of this invention. It is a flowchart for demonstrating the process sequence of control module 12-k at the time of the stream registration of the slave FEC server 1-k in the 2nd Embodiment of this invention.

Explanation of symbols

  1 FEC Server, 2 Internet, 3 Intranet, 4,7 Streaming Server, 5,6 Streaming Client, 11-1 to 11-n FEC Module, 12, 12-1 to 12-m Control Module, 13 Ethernet (registered trademark) Switch, 14 Internet port, 15 Intranet port, 21-1 to 21-n, 31 CPU, 22-1 to 22-n, 32 Memory.

Claims (10)

A server that connects a network with high packet loss and a network with low packet loss,
Receiving means for receiving a stream via the network with a low packet loss;
Encoding means for encoding the stream received by the receiving means into an error correctable format;
A transmission unit configured to transmit the stream encoded by the encoding unit via the network with a large packet loss. The server includes a plurality of the encoding means,
The server further includes a distribution unit for determining which stream the packet received by the reception unit belongs to, and distributing the packet to one of the plurality of encoding units according to the determination result. Item 1. The server according to item 1.   The server further determines whether or not the newly registered stream matches the registered stream based on a source address and a destination address of the newly registered stream, and if not, the plurality of encodings 3. The server according to claim 2, further comprising a control unit for determining to which encoding unit the newly registered stream is distributed according to a load state of the unit and instructing the distribution unit.   The control means newly registers based on the sum of the stream rates determined for each error correction parameter and the maximum stream rate for each error correction parameter that can be processed by one encoding means. 4. The server according to claim 3, wherein the server determines to which encoding means the stream is distributed. A server that connects a network with high packet loss and a network with low packet loss,
Receiving means for receiving a stream encoded in an error-correctable format via the packet loss network;
Decoding means for decoding the stream received by the receiving means;
A transmission unit configured to transmit the stream decoded by the decoding unit via the network with less packet loss. The server includes a plurality of the decoding means,
The server further includes a distribution unit for determining which stream the packet received by the reception unit belongs to, and distributing the packet to one of the plurality of decoding units according to the determination result. Item 6. The server according to item 5.   The server further determines whether or not the newly registered stream matches the registered stream based on a source address and a destination address of the newly registered stream, and if not matched, the plurality of decodes 7. The server according to claim 6, comprising control means for determining which decoding means to distribute the newly registered stream according to the load status of the means and instructing the distribution means.   The control means newly registers based on the sum of the stream rates determined for each error correction parameter and the maximum stream rate for each error correction parameter that can be processed by one decoding means. 8. The server according to claim 7, wherein the server determines to which decoding means the stream is distributed. A server that connects a network with a high packet loss and a network with a low packet loss, and includes a master server and a slave server,
The master server has a first receiving means for receiving a stream via the network with little packet loss or the adjacent slave server;
A plurality of first encoding means for encoding the stream received by the first receiving means into a format capable of error correction;
Determining which stream the packet received by the first receiving means belongs to, and allocating the packet to either the slave server or the plurality of first encoding means according to the determination result First distribution means;
First transmitting means for transmitting the stream encoded by the first plurality of encoding means via the network having a large packet loss or the server of the adjacent slave;
The slave server includes a second receiving means for receiving a stream via the network with less packet loss, the master server, or another adjacent slave server;
A second plurality of encoding means for encoding the stream received by the second receiving means into a format capable of error correction;
Second distribution means for determining which stream the packet received by the second reception means belongs to and distributing the packet to any of the plurality of second encoding means according to the determination result When,
And a second transmission means for transmitting the stream encoded by the second plurality of encoding means via the network with many packet losses, the master server or the adjacent slave server. A server that connects a network with a high packet loss and a network with a low packet loss and includes a master server and a slave server,
The master server has a first receiving means for receiving a stream encoded in an error-correctable format via the packet loss network or the adjacent slave server;
A first plurality of decoding means for decoding the stream received by the first receiving means;
Determining which stream the packet received by the first receiving means belongs to, and allocating the packet to either the slave server or the first plurality of decoding means according to the determination result First distribution means;
First transmission means for transmitting the stream decoded by the first plurality of decoding means via the network with little packet loss or the server of the adjacent slave,
The slave server has a second receiving means for receiving a stream encoded in an error-correctable format via the network with low packet loss, the master server, or another adjacent slave server;
A second plurality of decoding means for decoding the stream received by the second receiving means;
Second distribution means for determining which stream the packet received by the second reception means belongs to and distributing the packet to one of the plurality of second decoding means according to the determination result When,
And a second transmission means for transmitting the stream decoded by the second plurality of decoding means via the network with little packet loss, the master server, or the adjacent slave server.

Images